f5c62ee0fa
The insertion point of COPY isn't always optimal and could eventually lead to a worse block layout, see the regression test in the first commit. This change affects many architectures but the amount of total instructions in the test cases seems too be slightly lower.
83 lines
2.4 KiB
LLVM
83 lines
2.4 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 3
|
|
; RUN: llc -mtriple=arm--none-eabi < %s | FileCheck %s
|
|
|
|
%type_80 = type [20 x i32]
|
|
|
|
declare void @func_80(ptr, ptr byval(%type_80))
|
|
|
|
define void @test_80() {
|
|
; CHECK-LABEL: test_80:
|
|
; CHECK: @ %bb.0:
|
|
; CHECK-NEXT: .save {r11, lr}
|
|
; CHECK-NEXT: push {r11, lr}
|
|
; CHECK-NEXT: .pad #152
|
|
; CHECK-NEXT: sub sp, sp, #152
|
|
; CHECK-NEXT: add r0, sp, #72
|
|
; CHECK-NEXT: ldr r2, .LCPI0_0
|
|
; CHECK-NEXT: add r0, r0, #12
|
|
; CHECK-NEXT: mov r1, sp
|
|
; CHECK-NEXT: .LBB0_1: @ =>This Inner Loop Header: Depth=1
|
|
; CHECK-NEXT: ldr r3, [r0], #4
|
|
; CHECK-NEXT: subs r2, r2, #4
|
|
; CHECK-NEXT: str r3, [r1], #4
|
|
; CHECK-NEXT: bne .LBB0_1
|
|
; CHECK-NEXT: @ %bb.2:
|
|
; CHECK-NEXT: add r3, sp, #72
|
|
; CHECK-NEXT: mov r0, #0
|
|
; CHECK-NEXT: ldm r3, {r1, r2, r3}
|
|
; CHECK-NEXT: bl func_80
|
|
; CHECK-NEXT: add sp, sp, #152
|
|
; CHECK-NEXT: pop {r11, lr}
|
|
; CHECK-NEXT: mov pc, lr
|
|
; CHECK-NEXT: .p2align 2
|
|
; CHECK-NEXT: @ %bb.3:
|
|
; CHECK-NEXT: .LCPI0_0:
|
|
; CHECK-NEXT: .long 68 @ 0x44
|
|
%a = alloca %type_80
|
|
call void @func_80(ptr null, ptr %a)
|
|
ret void
|
|
}
|
|
|
|
%type_4000 = type [1000 x i32]
|
|
|
|
declare void @func_4000(ptr, ptr byval(%type_4000))
|
|
|
|
define void @test_4000() {
|
|
; CHECK-LABEL: test_4000:
|
|
; CHECK: @ %bb.0:
|
|
; CHECK-NEXT: .save {r11, lr}
|
|
; CHECK-NEXT: push {r11, lr}
|
|
; CHECK-NEXT: .pad #4000
|
|
; CHECK-NEXT: sub sp, sp, #4000
|
|
; CHECK-NEXT: sub sp, sp, #920
|
|
; CHECK-NEXT: sub sp, sp, #3072
|
|
; CHECK-NEXT: add lr, sp, #3072
|
|
; CHECK-NEXT: ldr r2, .LCPI1_0
|
|
; CHECK-NEXT: add r0, lr, #920
|
|
; CHECK-NEXT: mov r1, sp
|
|
; CHECK-NEXT: add r0, r0, #12
|
|
; CHECK-NEXT: .LBB1_1: @ =>This Inner Loop Header: Depth=1
|
|
; CHECK-NEXT: ldr r3, [r0], #4
|
|
; CHECK-NEXT: subs r2, r2, #4
|
|
; CHECK-NEXT: str r3, [r1], #4
|
|
; CHECK-NEXT: bne .LBB1_1
|
|
; CHECK-NEXT: @ %bb.2:
|
|
; CHECK-NEXT: ldr r1, [sp, #3992]
|
|
; CHECK-NEXT: mov r0, #0
|
|
; CHECK-NEXT: ldr r2, [sp, #3996]
|
|
; CHECK-NEXT: ldr r3, [sp, #4000]
|
|
; CHECK-NEXT: bl func_4000
|
|
; CHECK-NEXT: add sp, sp, #920
|
|
; CHECK-NEXT: add sp, sp, #3072
|
|
; CHECK-NEXT: add sp, sp, #4000
|
|
; CHECK-NEXT: pop {r11, lr}
|
|
; CHECK-NEXT: mov pc, lr
|
|
; CHECK-NEXT: .p2align 2
|
|
; CHECK-NEXT: @ %bb.3:
|
|
; CHECK-NEXT: .LCPI1_0:
|
|
; CHECK-NEXT: .long 3988 @ 0xf94
|
|
%a = alloca %type_4000
|
|
call void @func_4000(ptr null, ptr %a)
|
|
ret void
|
|
}
|